Mott Transition in the $A15$ Phase of ${\mathrm{Cs}}_3{\mathrm{C}}_{60}$: Absence of a Pseudogap and Charge Order

We present a detailed NMR study of the insulator-to-metal transition induced by an applied pressure $p$ in the $A15$ phase of ${\mathrm{Cs}}_3{\mathrm{C}}_{60}$. We evidence that the insulating antiferromagnetic (AFM) and superconducting (SC) phases coexist only in a narrow $p$ range. At fixed $p$, in the metallic state above the SC transition $T_c$, the ${}^{133}\mathrm{Cs}$ and ${}^{13}\mathrm{C}$ NMR spin-lattice relaxation data are seemingly governed by a pseudogaplike feature. We prove that this feature, also seen in the ${}^{133}\mathrm{Cs}$ NMR shift data, is rather a signature of the Mott transition which broadens and smears out progressively for increasing $(p,T)$. The analysis of the variation of the quadrupole splitting ${\nu }_Q$ of the ${}^{133}\mathrm{Cs}$ NMR spectrum precludes any cell symmetry change at the Mott transition and only monitors a weak variation of the lattice parameter. These results open an opportunity to consider theoretically the Mott transition in a multiorbital three-dimensional system well beyond its critical point.

Figure 1

(a) The ${}^{133}\mathrm{Cs}$ NMR intensity corrected for its Curie $1/T$ dependence in the spectral range delineated in Fig. S1(a) of Ref. [21]. For any $p$, the intensity loss due to the AFM contribution is seen to onset below 50 K. (b) AFM fraction obtained by integrating the intensity between 50 and 35 K. Comparison with the SC fractions obtained in Ref. [12] from diamagnetic data.

Figure 2

(a) $(T_{1}T{)}^{-1}$ for ${}^{13}\mathrm{C}$ and ${}^{133}\mathrm{Cs}$ taken at 5.9 kbar with the vertical axes scaled. (b) The ${}^{133}(T_{1}T{)}^{-1}$ data are plotted versus $T$ for a series of pressures $p$. The Curie-Weiss fits for 1 bar and 4.1 kbar (see the text) are plotted as solid and dotted lines, respectively. The step associated with the MIT shifts towards higher $T$ with increasing $p$.

Figure 3

The ${}^{133}\mathrm{Cs}$ NMR spectral analyses done in Sec. II of Ref. [21] yield (a) $K_{\text{iso}}$ on a large ($T$, $p$) range which displays step variations for increasing $T$ similar to those found for $(T_{1}T{)}^{-1}$ in Fig. 2 and (b) ${\nu }_Q$ data do reveal small jumps of $\sim 1\mathrm{kHz}$ for $5.4<p<6.4\mathrm{kbar}$ which disappear at higher $p$. Lines are guides to the eye.

Figure 4

MIT transition temperatures $T_Q$, $T_{\text{mid}(T_1)}$, $T_{\text{mid}(K)}$, and $T_{\max }$ deduced from ${}^{133}\mathrm{Cs}$ NMR data for $p>5.4\mathrm{kbar}$. The $T_Q$ data mark the weak first-order regime (solid line), while $T_{\max }$ delineates the apparent upper half width of the transition given as the blank zone. This suggests that the transition becomes a crossover (dotted line) at high $(T,p)$ beyond a critical point $p_c\simeq 7\mathrm{kbar}$ (see the text). The AFM and SC fractions measured in Fig. 1 yielded $p_{c0}=5.1\pm 0.3\mathrm{kbar}$ below 35 K. The phase diagram is completed by the horizontal $T_N$ line below 5.1 kbar, and the $T_c$ dome above 5.1 kbar.